fix(kimi): omit temperature entirely for Kimi/Moonshot models (#13157)

Kimi's gateway selects the correct temperature server-side based on the
active mode (thinking -> 1.0, non-thinking -> 0.6).  Sending any
temperature value — even the previously "correct" one — conflicts with
gateway-managed defaults.

Replaces the old approach of forcing specific temperature values (0.6
for non-thinking, 1.0 for thinking) with an OMIT_TEMPERATURE sentinel
that tells all call sites to strip the temperature key from API kwargs
entirely.

Changes:
- agent/auxiliary_client.py: OMIT_TEMPERATURE sentinel, _is_kimi_model()
  prefix check (covers all kimi-* models), _fixed_temperature_for_model()
  returns sentinel for kimi models.  _build_call_kwargs() strips temp.
- run_agent.py: _build_api_kwargs, flush_memories, and summary generation
  paths all handle the sentinel by popping/omitting temperature.
- trajectory_compressor.py: _effective_temperature_for_model returns None
  for kimi (sentinel mapped), direct client calls use kwargs dict to
  conditionally include temperature.
- mini_swe_runner.py: same sentinel handling via wrapper function.
- 6 test files updated: all 'forces temperature X' assertions replaced
  with 'temperature not in kwargs' assertions.

Net: -76 lines (171 added, 247 removed).
Inspired by PR #13137 (@kshitijk4poor).
This commit is contained in:
Teknium 2026-04-20 12:23:05 -07:00 committed by GitHub
parent c1977146ce
commit 3cba81ebed
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
10 changed files with 170 additions and 246 deletions

View file

@ -47,12 +47,19 @@ def _effective_temperature_for_model(
model: str,
base_url: Optional[str] = None,
) -> Optional[float]:
"""Return a fixed temperature for models with strict sampling contracts."""
"""Return a fixed temperature for models with strict sampling contracts.
Returns ``None`` when the model manages temperature server-side (Kimi);
callers must omit the ``temperature`` kwarg entirely in that case.
"""
try:
from agent.auxiliary_client import _fixed_temperature_for_model
from agent.auxiliary_client import _fixed_temperature_for_model, OMIT_TEMPERATURE
except Exception:
return None
return _fixed_temperature_for_model(model, base_url)
result = _fixed_temperature_for_model(model, base_url)
if result is OMIT_TEMPERATURE:
return None # caller must omit temperature
return result